One of the major ongoing projects of PLRP is classifying all the different kinds of microbialites in the lake. Although there are other places in the world where living microbialites exist, Pavilion Lake is distinctive for the diversity of the structures found there.

"Pavilion has, I think, greater morphological diversity than any other modern environment where people are studying microbialites. That's what makes it so special," said Rebekah Shepard, a PLRP team member and DeepWorker pilot who did her graduate research in how microbialites form. (Note: the word "morphology" comes up a lot when you talk to people doing microbialite research. It refers to the size and shape of something, its overall form.)

The first step in figuring out what's going on is data collection. In the case of microbialites, that means taking pictures. When PLRP pilots conduct their DeepWorker flights, they take video images of everything they see. By now, they have tens of thousands of video frames stored on hard drives, and this year they're capturing thousands more.

But having a bunch of hard drives full of microbialite photos doesn't do much good if you don't have a system for analyzing them. At first, people referred to various formations by comparing them to vegetables like artichokes and cauliflower. But a couple of years ago, it dawned on Shepard that a more systematic approach was needed to fully understand what was going on in the lake.

Modern microbialites like those at Pavilion Lake are rare, so no-one previously had developed a classification system for them. But there has been a lot of study of ancient stromatolites, the fossilized remains of ancient formations similar to microbialites.

So Shepard borrowed from a classification system used for the older stromatolites and adapted it for use in classifying the Pavilion Lake structures.

The modified system provides a multiple-choice approach to detailing the morphology of both the large mounds and the individual components that make up the larger constructs.

One question Shepard added, involving information not generally available when dealing with rocks billions of years old, was a choice about what type of surface, or substrate, each of the microbialites was growing on, such as sediment, algae, trees or rocks.

Still, someone had to go through all those video frames and painstakingly describe what they saw. The bulk of that work was done by Margarita Marinova, Zena Cardman, Mary Beth Wilhelm and Shepard herself, although many other PLRP participants have spent hours glued in front of computer screens trying to decide, for example, whether to describe a particular microbialite as "bulbous" or "turbinate."

Until recently, there were only a handful of computers on which the classification work could be done, because the system was built using MatLab, a math-and-science programming environment. This year, however, the job got a lot easier, thanks to Nick Wilkinson, who developed a Web-based version of the system with a graphical user interface that anyone, with a few minutes' training, can use.

Even me. Yes, I am now trained in the use of MAPPER (Morphology Analysis Project for Participatory Exploration and Research). Wilkinson patiently sat by my side while I slogged through some 300 frames of video. The only problem was that the video frames I was analyzing didn't contain any microbialites. Just mud. That made the classification process relatively simple, but truth be told, kind of boring.

Later in the day, Cardman took pity on me and showed me a few images that had microbialites in them, so I did get a little practice at describing them. I found myself scratching my head a lot, though. It's not always easy to tell the difference between bulbous and turbinate. Fortunately, there's an "Unsure" button one can click when the head-scratching gets too severe.

Because of this uncertainty factor, the plan is to have all the video reviewed not just once, but multiple times. To that end, Wilkinson is working on a version of the MAPPER software for use by the general public, something along the lines of SETI@home.

Participants would download a set of video frames off the Web, analyze them, and return the data to PLRP. (Friendly suggestion: edit out the shots of mud before you go public.)

Information entered into the classification system goes into a massive data base, which eventually will be combined with other data sets, such as the bathymetric map of the lake I wrote about in my last post.

Scientists will then be able to pose queries such as, Is there a strong correlation between the size of the mounds with the depth at which they occur? Being able to answer these kinds of questions are what the scientists working on PLRP hope ultimately will enable them to figure out how and why these amazing structures form.

"Will it work?" Shepard wondered out loud at the end of our conversation, just before she had to run off to pilot a sub. "No idea. It's never been done." She paused, smiled and then added, "I try not to mention this to the people doing the classifying."

del.icio.us	Digg	Reddit
YahooMyWeb	Google	Facebook

Memory Foam Mattress Review

SPACE MEDIA NETWORK PROMOTIONS Solar Energy Solutions Tempur-Pedic Mattress Comparison ... Newsletters :: SpaceDaily :: SpaceWar :: TerraDaily :: Energy Daily XML Feeds :: Space News :: Earth News :: War News :: Solar Energy News